GerdH˜onninger HalogenOxideStudiesintheBoundaryLayerbyMultiAxisDifierentialOpticalAbsorptionSpectroscopyandActiveLongpath-DOAS

Differential Optical Absorption Spectroscopy and Active Longpath-DOAS

Gerd H¨onninger

Combined Faculties for the Natural Sciences and for Mathematics of the Ruperto Carola University of Heidelberg, Germany

for the degree of Doctor of Natural Sciences

presented by

Diplom-Physicist: Gerd H¨onninger born in: Bad Mergentheim

Oral examination: 14.02.2002

Multi Axis

Differential Optical Absorption Spectroscopy and Active Longpath-DOAS

Referees: Prof. Dr. Ulrich Platt

Prof. Dr. Konrad Mauersberger

flusses auf Ozon von großer Bedeutung f¨ur die Chemie der Troposph¨are. Dies wurde zum ersten Mal w¨ahrend Episoden von v¨olligem Abbau des bodennahen Ozons nach Sonnen- aufgang im polaren Fr¨uhjahr durch Bromoxid (BrO) entdeckt. Halogenoxidvorkommen wurden zwischenzeitlich auch in mittleren Breiten an K¨usten (Jodoxid, IO und Joddioxid, OIO) und im Toten Meer Becken (BrO) gefunden. Neue Ergebnisse von Feldmessungen, die im Rahmen dieser Arbeit an Meßorten in der Arktis und in mittleren Breiten durch- gef¨uhrt wurden, werden hier vorgestellt. BrO und seine vertikale Verteilung in Boden- n¨ahe wurden w¨ahrend Ozonverlustereignissen in arktischen Gebieten Kanadas in hohen und mittleren geographischen Breiten beobachtet. Die ersten gleichzeitigen Messungen von BrO und IO in der arktischen bodennahen Grenzschicht wurden mit bodengest¨utzter MAX-DOAS gemacht, die im Rahmen dieser Arbeit entwickelt wurde. Die Messungen an der Hudson Bay stellen die s¨udlichsten und ersten direkten Messungen von bodenna- hem Ozonverlust in der Arktis durch BrO dar und erlaubten zum ersten Mal, die Tag- und Nachtchemie von Bromoxid zeitlich hochaufgel¨ost zu untersuchen. Molekulares Brom wurde als eine wichtige n¨achtliche Reservoirsubstanz identifiziert, die bei Sonnenaufgang nach ihrem photolytischen Abbau Ozonverluste startet. Die ersten gleichzeitigen, zeitlich hochaufgel¨osten Messungen von BrO und elementarem Quecksilber in der Gasphase in der kanadischen Arktis st¨utzen die vorgeschlagene Schl¨usselrolle von BrO als Oxidationsmit- tel von elementarem Quecksilber in der Gasphase w¨ahrend des polaren Fr¨uhjahrs. Weitere Messungen von Halogenoxiden an Reinluft- und m¨aßig verschmutzten Meßstationen erga- ben zus¨atzliche Konzentrationsdaten und Obergrenzen f¨ur Halogenoxide.

Halogen Oxide Studies in the Boundary Layer by Multi Axis Differential Optical Absorption Spectroscopy and Active Longpath DOAS

The importance of reactive halogen species (particularly halogen oxides) in the troposphere is due to their strong effect on tropospheric ozone levels, which has been first discovered during surface ozone depletion episodes in the polar boundary layer after polar sunrise (bromine oxide, BrO). Halogen oxides have also been reported from mid-latitude coastal sites (iodine oxide, IO and iodine dioxide, OIO) and from the Dead Sea basin (BrO).

Results from the field studies performed in framework of this thesis at Arctic and mid- latitude locations are presented here. BrO and its vertical profile near the ground has been observed during surface ozone destruction in the Canadian high and low Arctic. First simultaneous measurements of BrO and IO in the Arctic boundary layer were performed by ground-based MAX-DOAS, which was developed within this work. The measurements at the Hudson Bay represent the southernmost and first direct observations of Arctic surface ozone losses due to BrO and allowed for the first time to study the day and nighttime chemistry of BrO at high time resolution. Molecular bromine was inferred to be a major reservoir compound during the night, initiating sunrise ozone destruction upon photolysis. The first simultaneous measurements of BrO and gas phase elemental mercury at high time resolution in the Canadian Arctic support the proposed key role of BrO as oxidant for mercury in the gas phase during polar springtime. Further measurements of halogen oxides at clean and moderately polluted mid-latitude coastal sites yielded new field data on halogen oxide concentrations and upper limits at different pollution levels.

1 Introduction 1

2 The Atmospheric Chemistry of Halogens 5

2.1 Ozone in the Atmosphere . . . 5

2.1.1 Stratospheric Ozone . . . 7

2.1.2 Tropospheric Ozone . . . 10

2.2 Reactive Halogen Species in the Troposphere . . . 13

2.2.1 Reaction Pathways of RHS . . . 14

2.2.2 Sources of RHS . . . 22

2.2.3 Sinks of RHS . . . 26

2.3 Tropospheric Bromine Chemistry . . . 27

2.3.1 Sources and Reaction Cycles of Reactive Bromine . . . 31

2.3.2 The Lifetime of Bromine Oxide Radicals in the Boundary Layer . . 34

2.3.3 Sinks of Reactive Bromine . . . 36

2.4 Tropospheric Iodine Chemistry . . . 36

2.4.1 Sources and Reaction cycles of Reactive Iodine . . . 36

2.4.2 Fate of Reactive Iodine . . . 37

2.5 Mercury in the Atmosphere . . . 39

2.5.1 Sources and Partitioning of Mercury in the Atmosphere . . . 40

2.5.2 Sinks of Atmospheric Mercury and the Role of Halogens . . . 43

3 From DOAS to MAX-DOAS 47 3.1 DOAS Overview . . . 47

3.1.1 The Measurement Principle . . . 48

3.1.2 The Analysis Procedure . . . 52

3.1.3 Corrections to the Measured Spectra . . . 54

3.1.4 Error Estimation . . . 56

3.1.5 The Effects of Residual Structures . . . 56

3.1.6 Differential Cross Sections . . . 57

3.1.7 Example for a DOAS Evaluation . . . 57 i

3.2.2 Fraunhofer Structures . . . 61

3.2.3 The Ring Effect . . . 62

3.2.4 The Solar I₀ Effect . . . 64

3.3 Scattering Processes in the Atmosphere . . . 65

3.3.1 Rayleigh Scattering . . . 65

3.3.2 Raman Scattering . . . 66

3.3.3 Mie Scattering . . . 67

3.4 Radiative Transfer in the Atmosphere . . . 69

3.4.1 The Airmass Factor Concept . . . 69

3.4.2 An Improved Airmass Factor Concept . . . 72

3.5 Multi-Axis-DOAS . . . 77

3.5.1 Off Axis DOAS . . . 78

3.5.2 Multi-Axis-DOAS Observations from the Ground . . . 80

3.5.3 MAX-DOAS Airmass Factors . . . 84

3.5.4 Airborne MAX-DOAS (AMAX-DOAS) . . . 90

3.5.5 Multi-Axis-DOAS Instruments . . . 91

3.5.6 Other Possible MAX-DOAS Applications . . . 94

4 Instrumental Setups 99 4.1 The MAX-DOAS Instrument . . . 99

4.1.1 Entrance Optics . . . 99

4.1.2 Spectrograph and Detector Unit . . . 100

4.2 The Active Long Path-DOAS System . . . 102

4.2.1 The LP-DOAS Telescope . . . 102

4.2.2 The Light Source . . . 103

4.3 New Miniature DOAS Instruments . . . 104

4.3.1 The Mini-MAX-DOAS System . . . 104

4.3.2 The Portable Active LP-DOAS System . . . 109

5 Field Measurements 111 5.1 Preparatory Field Studies . . . 111

5.1.1 Zugspitze MAX-DOAS Measurements . . . 111

5.1.2 Indian Ocean Field Study . . . 120

5.2 ALERT2000 Field Study . . . 127

5.2.1 Measurement Sites . . . 129

5.2.2 Meteorological Parameters . . . 130

5.2.3 MAX-DOAS BrO Measurements . . . 132

5.2.4 MAX-DOAS IO Measurements . . . 138

5.3.1 Measurement Site . . . 143

5.3.2 Meteorological Parameters . . . 144

5.3.3 Active LP-DOAS BrO, IO, OIO Measurements . . . 144

5.3.4 Nitrogen Compounds at Finokalia . . . 147

5.4 Hudson Bay Campaign . . . 149

5.4.1 Measurement Site . . . 149

5.4.2 Meteorological Parameters . . . 149

5.4.3 First LP-DOAS Measurements at the Hudson Bay . . . 151

5.4.4 Etalon Structures . . . 156

5.4.5 Local Pollution . . . 160

5.4.6 MAX-DOAS BrO Measurements . . . 161

5.4.7 Ozone Measurements at Kuujjuarapik . . . 163

5.4.8 Mercury Measurements at Kuujjuarapik . . . 166

6 Results 167 6.1 BrO in the Free Troposphere . . . 167

6.2 Halogen Oxides in the Southern Indian Ocean . . . 167

6.3 Results from ALERT2000 . . . 168

6.3.1 Time Series of BrO during ALERT2000 . . . 168

6.3.2 The Vertical Extent of the BrO Layer . . . 170

6.3.3 BrO during Ozone Depletion . . . 171

6.3.4 Comparison with GOME Vertical Column Densities . . . 180

6.3.5 Iodine Chemistry in the Arctic Boundary Layer . . . 182

6.4 Results from Crete2000 . . . 183

6.4.1 Upper Limits of Halogen Oxides in the Mediterranean Region . . . . 183

6.4.2 Results from MOCCA Model Simulation . . . 184

6.5 Results from the Hudson Bay Measurements . . . 187

6.5.1 First Halogen Oxide Measurements at Hudson Bay . . . 188

6.5.2 Day/Nighttime Chemistry of Bromine Oxide . . . 190

6.5.3 Upper Limits of the Halogen Oxides IO, OIO, OBrO and OClO . . . 197

6.5.4 Comparison of LP-DOAS and MAX-DOAS Results . . . 198

6.5.5 Comparison of Boundary Layer BrO Data with GOME Maps . . . . 200

6.5.6 Model Results of Day/Nighttime Chemistry at Kuujjuarapik . . . . 204

6.6 The Role of BrO as Oxidant for Gas Phase Mercury . . . 207

6.6.1 Reactive Bromine - Mercury - Interaction . . . 209

7 Summary and Outlook 211

A The Etalon-Effect 215

B UFS Zugspitze Data Winter 1999/2000 217

C Data from Marion Dufresne, December 2000 223 D 5 day Back Trajectories Overview for ALERT2000 225

E MOCCA Simulation Results for Crete2000 227

List of Figures 239

List of Tables 241

References 260

Acknowledgements 262

Introduction

Atmospheric ozone has become a major concern since the discovery of the ozone hole by Farman et al.[1985]. This observation proved, to what extent anthropogenic activ- ities can influence environmental conditions in the atmosphere. At that time total at- mospheric ozone above Antarctica during spring was found to have decreased to below 200 Dobson Units¹, which was only about 70% of the values observed in the years be- fore. In the following years springtime ozone columns fell even below 100 DU. This ob- servation was very important, since the ozone layer protects life on Earth against the harmful ultraviolet radiation of the sun. A slight decrease in the stratospheric ozone concentration could be explained by the suggestions of Molina and Rowland[1974] and Stolarski and Cicerone [1974] that reactive chlorine compounds could be involved in cat- alytic ozone destruction cycles. The predicted global ozone loss due to anthropogenic emis- sions of CFCs (chlorofluorocarbons) and halons (brominated organic compounds) into the atmosphere was estimated to 10-20% over the next 50-100 years. However, they could not explain the observation of the Antarctic ozone hole reported by Farman et al.[1985]. In 1986, Solomon et al. suggested that chlorine compounds might react on the surfaces of polar stratospheric clouds (PSCs) which occur at the low temperatures in the Antarctic stratosphere during polar night. McElroy et al.[1986] proposed additional ozone destruc- tion cycles involving combined chemistry of reactive chlorine and bromine. These theories were confirmed by many studies in the following years, which proved the key role of reac- tive halogen compounds (chlorine and bromine) in stratospheric ozone chemistry.

In the mid 1980s sudden depletion events of ozone in the planetary boundary layer² were reported during springtime from several Arctic sites [Oltmans and Komhyr 1986;

1One Dobson Unit (DU) corresponds to an O3 column of 0.01 mm at standard pressure and temperature

2The Planetary Boundary Layer (PBL) is the lowermost region of the troposphere which is directly influenced by friction on the earth’s surface. Vertical mixing of trace gases and momentum is usually fast, leading to a generally well-mixed boundary layer of about 1 km vertical extent, depending on the stability conditions.

1

Bottenheim et al. 1986]. Within hours to days ozone levels at the surface frequently dropped to unmeasurable values in the weeks and months following polar sunrise. It turned out that also in the planetary boundary layer reactive halogen compounds (mainly bromine) are involved in catalytic ozone destruction during Arctic spring [Barrie et al. 1988;Barrie and Platt1997]. Most recently, indications and direct measure- ments of the presence of reactive bromine in the free troposphere have been reported as well [Harder et al.1998; Frieß et al. 1999; McElroy et al. 1999; van Roozendael et al. 2000;

Fitzenberger et al. 2000]. Halogen species are therefore assumed to have an influence on the ozone chemistry of the atmosphere on a global scale. Ground-based mea- surements have subsequently discovered the reactive halogen species bromine monox- ide [Hausmann and Platt 1994], iodine monoxide [Alicke et al.1999] and iodine dioxide [Hebestreit2001; Allan et al.2001] by Longpath-DOAS measurements in the boundary layer. However, many open questions still remained:

• What is the global distribution of reactive halogen species in the boundary layer?

• What are the release processes for the reactive halogen compounds observed at different locations?

• What are the levels of reactive halogens in the free troposphere?

• How can reactive halogen species influence the ozone budget in the troposphere on a global scale?

• What are the consequences for the oxidizing capacity of the atmosphere and the global radiation budget?

In this PhD thesis field studies on reactive halogen species were carried out at various coastal sites yielding highly interesting results which have not been reported in the liter- ature so far.

Outline of the thesis

First an overview of the relevant atmospheric chemistry of ozone and reactive halogen species is given in the second chapter. Both reactive bromine and iodine chemistry is described in detail. An additional section is included on atmospheric mercury, which is supposed to be strongly influenced by reactive bromine in polar regions. Thethird chapter introduces the measurement technique and related concepts used in this study. The central technique applied in this work is the differential optical absorption spectroscopy (DOAS) method. An overview is given on the principles and details of the DOAS analysis procedure.

Special aspects of active and passive DOAS instruments, which were both applied in this work are also described. For passive DOAS of scattered sunlight the understanding of the processes which determine the radiative transfer in the atmosphere is essential. Therefore

the basic atmospheric processes and the airmass factor concept used for modelling the radiative transfer is described in this chapter. After a brief overview of the off axis DOAS method, the MAX-DOAS³technique is introduced, which was developed in this work based on off axis DOAS. Since MAX-DOAS represents a significant new development and opens a wide field of future applications, the method is described in detail and possible applications are discussed. The fourth chapter presents the employed hardware for the measurements.

The custom built MAX-DOAS system as well as a state of the art Longpath-DOAS system are described and the main components are characterized. Additionally, the new mini- DOAS instruments tested during this thesis are briefly shown. The fifth chapter presents two preparatory and the three main field campaigns carried out as major experimental part of this work. The respective measurement location and climatology as well as the performed DOAS and related measurements are characterized in detail. Finally the results of the field measurements of halogen oxide radicals are presented in the sixth chapter, together with interpretation of the data and comparison with previous findings and model studies, which have been performed in this work to simulate the situations encountered during the field studies. The main results of this work are summarized in the final chapter seven and a short outlook is given on the future of this research field.

3Multi-Axis-DOAS

The Atmospheric Chemistry of Halogens

In this chapter the atmospheric chemistry of halogens species, their impact on ozone chemistry, relevant radicals and their role as oxidants in the atmosphere is discussed.

A brief introduction on atmospheric ozone is given in section 2.1. An overview of the main reaction pathways of halogen species in the atmosphere and in particular in the troposphere is given in section 2.2. The sources and sinks of bromine compounds in the troposphere and their impact on tropospheric ozone, particularly in the boundary layer, is subject of section 2.3. Section 2.4 deals with the sources, recycling and sinks of iodine compounds in the marine boundary layer. The possible involvement of reactive halogens in the atmospheric cycling of mercury species is addressed in section 2.5.

2.1 Ozone in the Atmosphere

Ozone was first proposed as an atmospheric constituent by [Sch¨onbein1840]. Its existence in the troposphere was then established in 1858 by chemical means [Houzeau1858]. In the late 19^th century subsequent spectroscopic studies in the visible and ultraviolet regions showed, that ozone is present at a higher mixing ratio in the upper atmosphere than near the ground [Hartley 1881]. The presence of ozone in the stratosphere, with a maximum concentration between 15 and 30 km altitude, the so-called ozone layer, protects life from harmful UV radiation, which can affect the health of humans, animals and plants. In 1930 the first theory on the photochemical formation of ozone in the stratosphere predicting a maximum concentration around 20 km was proposed byChapman [1930] (see Figure 2.2).

The much lower actual ozone levels measured were subsequently explained by numerous chemical species present in the stratosphere, such as hydrogen and nitrogen compounds [Bates and Nicolet 1950; Crutzen 1970; Johnston 1971]. A possible role of halogen com- pounds in stratosphere was already proposed by Crutzen [1973]. In 1985 the discovery of

5

0 1 2 3 4 5 6 7 0

5 10 15 20 25 30 35

Altitude[km]

O₃ concentration [x10¹²cm^-3]

0.01 0.1 1 10

0 5 10 15 20 25 30 35

Tropopause

Altitude[km]

O₃ mixing ratio [ppm]

Figure 2.1: Typical ozone sonde profile from the ground to 35 km measured during the ALERT2000 field campaign

theAntarctic ozone holebyFarman et al.[1985] lead to a growing interest in stratospheric ozone chemistry.

Only about 10% of the total ozone column is located in the troposphere. However, also in the troposphere ozone is a key component being the most important precursor of hydroxyl radicals (OH). The OH radical is the most important oxidizing species in the daytime atmosphere (see e.g. review by Crutzen and Zimmermann[1991]) and therefore the key component in the degradation and removal of pollutants from the atmosphere. It is also the central compound in the formation of ozone in both, polluted and clean areas and contributes to the radiative forcing as a greenhouse gas.

Figure 2.1 shows a typical ozone vertical profile with the concentration and mixing ratio of ozone in the atmosphere as a function of altitude. The ozone sonde profile was measured during the ALERT2000 field campaign and represents typical background levels. In the following two sections the sources and sinks of stratospheric and tropospheric ozone will be explained.

Figure 2.2: The Chapman cycle

2.1.1 Stratospheric Ozone

Since ozone is a strong absorber in the ultraviolet spectral region (see cross section in Figure 3.2), it is one of the key species in the earth’s atmosphere. The formation and destruction of ozone in the stratosphere can be described by the so-called odd-oxygen chemistry. The production of ozone is initiated by the photolysis of molecular oxygen [Chapman1930]:

O₂+hν −→2O(³P) λ≤242nm (2.1a)

O(³P) +O2+M −→O3+M (2.1b)

Ozone is formed via the reaction (2.1b) of O + O₂ with a collision partner M. The following reactions lead to destruction of ozone:

O₃+hν −→O₂+O(¹D) λ≤320nm (2.2a)

O(¹D) +M −→O(³P) +M (2.2b)

O₃+hν −→O₂+O(³P) λ≤1180nm (2.2c)

2O(³P) +M −→O₂+M (2.2d)

O(³P) +O3−→2O2 (2.2e)

Reactions (2.1b) and the photolysis of ozone ((2.2a) and (2.2c)) rapidly interconvert O and O₃, which provides the rationale for the concept of odd oxygen (O and O₃). Even oxygen is defined as O₂. Since reaction (2.2d) is known to be too slow for it to play a part in stratospheric chemistry, (2.2e) represents the only loss process for odd oxygen in the chapman cycle. The reaction scheme is shown in Figure 2.2.

Soon it became clear that the observed ozone profiles cannot be explained by the Chapman cycle alone, but that other ozone destroying mechanisms must exist. Model calculations including only oxygen chemistry strongly overestimated the stratospheric ozone abundance by more than a factor of two. Therefore additional sinks for ozone must be important. A

set of ozone destroying reaction cycles, involving hydrogen oxides, was first proposed by Bates and Nicolet [1950]:

Cycle a)

O+OH−→O₂+H (2.3a)

H+O₂+M −→HO₂+M (2.3b)

O₃+HO₂−→2O₂+OH (2.3c)

net: O+O₃−→O₂ Cycle b)

O+OH −→O2+H (2.4a)

H+O₃+M −→OH+O₂+M (2.4b)

net: O+O₃ −→2O₂ Cycle c)

O3+OH−→O2+HO2 (2.5a)

O+HO₂−→OH+O₂ (2.5b)

net: O+O₃−→2O₂

OH is produced in the stratosphere by the reaction of oxygen atoms with water vapor or methane (CH4). The stratosphere appears to be very dry since the tropopause acts as a cold trap and therefore prevents tropospheric water vapor from mixing into the stratosphere.

Instead, water vapor is produced in the stratosphere by the following reactions:

CH₄+O(¹D)−→OH+CH₃ (2.6a)

CH4+OH −→H2O+CH3 (2.6b)

O(¹D) +H₂O −→2OH (2.6c)

The reaction chains (2.3) to (2.5) were the first of numerous catalytic reaction cycles proposed for the destruction of ozone: the compound responsible for the conversion of ozone to molecular oxygen (OH) is not consumed during the reaction cycles but only acts as a catalyst. It therefore remains available for the destruction of ozone unless it is removed by other sinks. Catalytic species have a strong impact on the ozone budget even at very low concentrations. The class of odd hydrogen compounds, i.e. OH and HO₂, can be summarized as HO_x.

Similar catalytic ozone destruction cycles involve nitrogen oxides [Crutzen 1970;

Johnston 1971]:

O₃+N O−→O₂+N O₂ (2.7a)

O+N O2−→N O+O2 (2.7b)

net: O+O3−→2O2

O₃+N O−→O₂+N O₂ (2.8a)

N O₂+O₃ −→N O₃+O₂ (2.8b)

N O₃+hν −→N O+O₂ (2.8c)

net: 2O₃+hν −→3O₂

Nitrogen oxides are produced in the stratosphere mainly by oxidation of N2O, which is a stable, well-mixed constituent of the atmosphere with a mixing ratio of ≈306 ppb. It is biogenically emitted from soils or the oceans and destroyed in the stratosphere:

N₂O+O(¹D)→2N O (2.9)

As for hydrogen compounds NO_x denotes the family of the reactive nitrogen compounds NO, NO₂ and NO₃, while all inorganic nitrogen species are summarized as NO_y.

Together with the following catalytic cycles involving chlorine and combined chlorine- bromine chemistry, the consideration of the catalytic cycles described above leads to an ozone budget in the stratosphere which is consistent with the observations [Molina and Rowland1974; Molina and Molina1987; McElroy et al. 1986].

Cl+O₃ −→ClO+O₂ (2.10a)

ClO+O −→Cl+O2 (2.10b)

net: O+O3 −→2O2

Cl+O3−→ClO+O2 (2.11a)

Cl+O₃−→ClO+O₂ (2.11b)

ClO+ClO+M −→Cl₂O₂+M (2.11c)

Cl₂O₂+hν −→Cl+ClO₂ (2.11d)

ClO₂+M −→Cl+O₂+M (2.11e)

net: 2O₃−→3O₂

Cl+O₃−→ClO+O₂ (2.12a)

Br+O₃−→BrO+O₂ (2.12b)

BrO+ClO−→Br+ClO2 (2.12c)

ClO₂+M −→Cl+O₂+M (2.12d)

net: 2O₃−→3O₂

Under normal conditions most of the chlorine in the stratosphere is usually bound up in the reservoir species hydrogen chloride (HCl) and chlorine nitrate (ClONO2) formed in the reactions

Cl+CH4 −→ HCl+CH3 (2.13)

ClO+N O₂+M −→ ClON O₂+M (2.14) Under the special conditions in the polar winter stratosphere (darkness, cold temperatures, stable polar vortex), heterogeneous reactions on the surface ofPolar Stratospheric Clouds (PSCs) can transform a huge fraction of chlorine into activated forms, e.g. into Cl₂ by the reaction:

ClON O2+HCl−→^het Cl2+HN O3 (2.15) As a result during polar winter molecular chlorine can accumulate inside the polar vortex.

As soon as sunlight is present the photolysis of Cl₂

Cl₂+hν −→Cl+Cl (2.16)

produces chlorine atoms which can efficiently destroy ozone in the catalytic cycles 2.10, 2.11 and - if bromine is present - 2.12. During polar spring these processes lead to the Stratospheric ozone hole, which was first discovered by Farman et al.[1985] over Antarc- tica.

2.1.2 Tropospheric Ozone

The production of ozone by the reaction sequence 2.1 is not possible in the troposphere, since UV light below 240 nm necessary to photolyze O₂ cannot penetrate into the tropo- sphere due to complete absorption by O2 (λ <242nm) and O3 (240nm < λ < 290nm) in the stratosphere. Therefore, except for the urban areas during pollution episodes, it was commonly assumed until the late 1970s that tropospheric ozone has its origin in the strato- sphere (e.g. [Junge 1963]). It was believed that stratospheric ozone was mixed through the tropopause region exhibiting a gradient towards the earth’s surface which was thought to

be the dominant sink. Fishman and Crutzen [1978] compared tropospheric ozone concen- trations in the two hemispheres with the then known source and sink mechanisms. It turned out that only about 50% of the tropospheric ozone has its origin in the stratosphere and is transported through the tropopause. Instead, as proposed byFishman and Crutzen [1978], the production of ozone in the troposphere is driven by NO_x chemistry and reactions in- volving methane or higher hydrocarbons. A key sequence for the formation and destruction of tropospheric ozone are reactions involving NO_x:

N O₂+hν −→N O+O(³P) λ≤420nm (2.17a) O(³P) +O₂+M −→O₃+M k₁ = 1.5·10⁻¹⁴ (2.17b) N O+O3−→N O2+O2 k2 = 1.8·10⁻¹² (2.17c) with the rate constants given in units of cm³ molec⁻¹ s⁻¹. The ozone concentration is then determined by the photo stationary steady state of reactions (2.17), which can be expressed as the so-calledLeighton ratio L:

L≡ [N O]

[N O2] = J(N O₂) [O3]·k2

(2.18) Remote regions are, in contrast to areas with anthropogenic influence, generally charac- terized by low NO_x with mixing ratios as low as 5-10 ppt e.g. in the Antarctic boundary layer [Jones et al.1999]. Similarly, in the arctic boundary layer NO_x mixing ratios typi- cally range between 10 and 20 ppt ([Beine et al.2001], see also section 5.4). Under these conditions, nitrogen and hydrogen oxides are able to produce ozone during the degradation of methane or higher hydrocarbons in the following reaction sequence:

CH₄+OH−→CH₃+H₂O (2.19a)

CH3+O2+M −→CH3O2+M (2.19b)

CH₃O₂+N O−→CH₃O+N O₂ (2.19c)

CH₃O+O₂−→CH₂O+HO₂ (2.19d)

CH₂O+hν −→^50% HCO+H (2.19e)

CH₂O+hν −→^50% H₂+CO (2.19f)

HCO+O2+M −→CO+HO2+M (2.19g)

H+O₂+M −→HO₂+M (2.19h)

HO₂+N O−→OH+N O₂ (2.19i)

N O₂+hν −→N O+O (2.19j)

O+O₂+M −→O₃+M (2.19k)

net: CH₄+xO₂−→CO+ 2H₂O+yO₃

The number of ozone molecules produced by the degradation of one methane molecule is y≈2.5. Ozone producing reaction schemes similar to (2.19) can also involve higher hydro- carbons. Besides the production of ozone, this reaction mechanism is an important source of carbon monoxide (CO) in the troposphere.

Carbon monoxide can produce additional ozone according to the following reactions [Fishman and Crutzen 1978]:

CO+OH −→CO₂+H (2.20a)

H+O₂+M −→HO₂+M (2.20b)

HO₂+N O −→OH+N O₂ (2.20c)

N O2+hν −→N O+O (2.20d)

O₂+O+M −→O₃+M (2.20e)

net: CO+ 2O₂ −→CO₂+O₃

Carbon monoxide, like CH₄, however, only produces ozone if the NO_x concentrations are above a threshold value. This is not the case for low NO_x conditions of remote areas with background conditions of typically 10 ppt. Instead, CO and hydrocarbons lead to destruction of ozone under low NO_x conditions:

CO+OH −→CO₂+H (2.21a)

H+O₂+M −→HO₂+M (2.21b)

HO2+O3 −→OH+ 2O2 (2.21c)

net: CO+O3 −→CO2+O2

The ozone mixing ratios in the unpolluted marine boundary (e.g. in the southern Indian ocean) range from ≈13 ppb during summer to≈30 ppb during winter [Gros et al.1998].

This pronounced seasonal variation of ozone is partly due to the seasonal variation of O₃ input from the stratosphere and by long range transport of ozone producing pollutants from biomass burning. The summer minimum of ozone can be explained by photochemical O₃ depletion which is controlled by the availability of OH radicals. The concentration of OH depends on the solar flux since OH is generated by the photolysis of O₃ to O(¹D), followed by its reaction with water vapor:

O₃+hν −→O₂+O(¹D) λ≤320nm (2.22a)

O(¹D) +H₂O−→2OH (2.22b)

The removal of OH occurs via the reaction with CO (reactions (2.20a) and (2.21a)), with CH₄ (reactions (2.19a) and (2.19b)) or reactions with other hydrocarbons.

The classical picture of the O₃ chemistry in the troposphere is, however, incomplete and

far from being completely understood.

In the weeks and months after polar sunrise, the ozone budget in the polar marine bound- ary layer is strongly affected by halogen chemistry. The mechanisms leading to the some- times complete depletion of surface ozone during such periods, thepolar tropospheric ozone hole, are described in the next section and the role of bromine is explained in section 2.3.

There are also several field studies at mid-latitudes showing a behavior of ozone mixing ratios which could not be explained by standard OH and NO_x chemistry. Recent measure- ments by Nagao et al.[1999] and Dickerson et al. [1999] suggest that ozone destruction by halogen chemistry could play a significant role in the remote marine boundary layer at mid- and low latitudes. Nagao et al. [1999] proposed that the observedsunrise ozone de- struction may be due to reactive halogens released from nighttime reservoir species which are rapidly photolyzed during sunrise.

2.2 Reactive Halogen Species in the Troposphere

Reactive Halogen Species (RHS) comprise the halogen atoms X, their monoxides XO, higher oxides X_nO_m, the hypohalous acids HOX, the halogen molecules X₂ and interhalo- gen compounds XY (X,Y = F, Cl, Br, I). In contrast to RHS the reactivity of reservoir species like halogen-NO_x compounds (XNO_x) or hydrogen halides (HX) are comparably slow. As will be described in this section, there are two main catalytic reaction cycles involving halogens which can destroy ozone in the troposphere, particularly in the marine boundary layer: Cycle (I) is based on the XO-self- or XO-YO-cross-reaction, cycle (II) on the reaction of XO with HO_x radicals.

Over the past decade significant amounts of XO were found in the marine bound- ary layer (MBL) of various coastal areas. Strong and sudden increases in the BrO mixing ratio during spring were found both in the Arctic [Hausmann and Platt 1994;

Tuckermann et al. 1997; Martinez et al.1999; McElroy et al. 1999] and in the Antarctic [Kreher et al.1997; Frieß1997; Frieß2001] boundary layer. Recently a study suggesting a free tropospheric background of 1-3 ppt of BrO based on multi-platform observations of BrO has been presented by van Roozendael et al. [2000]. Huge clouds of highly el- evated BrO amounts over the polar sea ice of both hemispheres, with areas spanning several thousand square kilometers, were observed from satellite [Wagner and Platt 1998;

Richter et al. 1998; Hegels et al.1998]. It has been proposed that these boundary layer BrO clouds may also contribute to BrO in the free troposphere [McElroy et al. 1999;

Roscoe et al. 2001]. These events of highly elevated BrO in polar regions, ground-based measurements showed levels of up to 30 ppt, were always coincident with the destruction of ozone in the boundary layer, indicating that reactive bromine is responsible for catalytic ozone destruction. Enhanced BrO in the boundary layer associated with ozone destruction was also detected over the Caspian Sea [Wagner et al. 2001] and in the Dead Sea basin

[Hebestreit et al.1999].

Iodine oxide (IO) was first observed at Mace Head, Ireland by Alicke et al.[1999] at levels of up to 6 ppt. Recently IO and also OIO were found at various coastal sites with mixing ratios in the ppt range: on Tenerife, Canary Islands and Cape Grim, Tas- mania [Allan et al.2000; Allan et al.2001], at Mace Head, Ireland [Allan et al.2000;

Hebestreit 2001], in the European Arctic [Wittrock et al. 2000] and in Antarctica [Frieß et al.2001; Frieß2001]. Figure 2.3 shows the regions where tropospheric reactive halogen species were measured. Besides DOAS measurements of halogen oxides photolyz- able bromine species (mostly HOBr, BrO) have been detected byImpey et al.[1999] using a Photolyzable Halogen Detector (PHD, conversion of reactive halogens to chloroacetone and bromoacetone and subsequent GC analysis). Hydrocarbon Clock measurements to derive chlorine and bromine atom concentrations have been reported from various lo- cations, which are not all shown on the map [Jobson et al.1994; Ramacher et al.1997;

Ramacher et al.1999; Solberg et al.1996; Ariya et al.1999]. Additionally, chemical am- plifier measurements by Perner et al.[1999] suggest that ClO is present in the Arctic boundary layer during ozone depletion periods.

2.2.1 Reaction Pathways of Reactive Halogen Species in the Troposphere

The main reaction schemes of the halogens Cl, Br and I are very similar regarding tropo- spheric chemistry¹. As will be discussed later, there are differences in the rate constants and different quantum yields concerning their photochemical reaction channels (see e.g.

Table 2.1). Therefore, if the reactions given are similar for the different halogens involved, X and Y will be used instead of the chemical symbols Cl, Br or I. Several rate constants and photolysis frequencies of RHS, which are important with regard to the two different ozone destruction cycles involving halogens, are listed in Table 2.1. Halogen atoms (X, Y) and their monoxides (XO, YO) are the key species in the ozone destruction cycles [Hausmann and Platt1994; LeBras and Platt 1995; Platt and Janssen1995]. Halogen ox- ides are formed in reaction with ozone (see reactions 2.23a and 2.23b below). Halogen atoms in the troposphere have a very short lifetime in the troposphere against their reac- tion with ozone (lifetime τ=0.08 s, 0.8 s, 0.8 s for Cl, Br, I). The typical reaction scheme for the first catalytic ozone destruction cycle involving reactive halogen species is

1As will be discussed at the end of this section, fluorine atoms mainly react with H2O or hydrocarbons to HF, which is rapidly removed from the atmosphere.

CTIVEHALOGENSPECIESINTHETROPOSPHERE15

"!$#&%'$!$#&% "!$#&% "!$#&% "!$#&%

"!$#&%'$!$#&% "!$#&%'$!$#&%

()*

+,

()*

+,

()* +,()*

+,

()* +,()*

+ ,

()*

+ ,

()*

+ ,

()*

+, - ./

- ./

- ./

- ./

- ./

- ./

- ./

- ./

- ./ 0 1 2

0 1 2

0 1 2

0

1 2

0

1 2

0

1 2

0

1 2

0

1 2

0

1 2

34 5

34 5

34 5

34 5

34 5

34 5

34 5

34 5

34 5

678

9 :

678

9 :

678

9 :

678

9 :

678

9 :

678

9 :

678

9 :

678

9 :

678

9 : ;<=

> ?

;<=

> ?

;<=

> ?

;<=

> ?

;<=

> ?

;<=

>?

;<=

>?

;<=

>?

;<=

> ? @AB

C D

@AB

C D

@AB

C D

@AB

C D

@AB

C D

@AB

CD

@AB

CD

@AB

CD

@AB

C D EFG

H I

EFG

H I

EFG

H IEFG

H I

EFG

H I

EFG

HI

EFG

HI

EFG

HI

EFG

H I JKL

M N

JKL

M N

JKL

M NJKL

M N

JKL

M N

JKL

MN

JKL

MN

JKL

MN

JKL

M N OP

QROP

QROP

QROP QROP

QROP Q ROP

Q ROP

Q R

OP

QR ST

UVST

UVST

UVST UVST

UVST U VST

U VST

U V

ST

UV WX

YZWX

YZWX

YZWX YZWX

YZWX Y ZWX

Y ZWX

Y Z

WX

YZ [\

]^[\

]^[\

]^[\

]^[\

]^[\

] ^[\

] ^[\

] ^

[\

]^ _`

ab_`

ab_`

ab_`

ab_`

ab_`

a b_`

a b_`

a b

_`

ab cd

efcd

efcd

efcd efcd

efcd e fcd

e fcd

e f

cd

ef g$hgihg$hg$hg$h

g$hgihg$hgih jlkim&nj'k$monjlkim&njlk

m&njlkm&n

jlkim&nj'k$monjlkim&nj'k

mon

p"qir&sp$q$rosp"qir&sp

qr&spqr&s

p"qir&sp$q$rosp"qir&sp

qros tvuiwyxtzu$w&xtvuiwyxtuwxtuwx tvuiwyxtzu$w&xtvuiwyxt

uwx {i|i}&~{v|$}o~{i|i}&~{i|}&~{i|}&~{|i}&~{|$}o~{|i}&~

{v|}o~

"€i&‚$€$o‚"€i&‚"€

&‚"€&‚"€i&‚$€$o‚"€i&‚$€o‚

ƒ"„i…&†ƒ$„$…o†ƒ"„i…&†ƒ"„

…†ƒ"„…†ƒ"„i…&†ƒ$„$…o†ƒ"„i…&†ƒ$„…† ‡'ˆ$‰$Šo‹‡ ˆ"‰$Šo‹‡'ˆ$‰$Šo‹‡'ˆ

‰$Š‹‡'ˆ

‰$Š‹‡'ˆ$‰$Šo‹‡ ˆ"‰$Šo‹‡'ˆ$‰$Šo‹‡ ˆ

‰$Š‹ ŒŽv"o‘Œli"o‘ŒŽv"o‘ŒŽ"

‘ŒŽ"

‘

ŒŽv"o‘Œli"o‘ŒŽv"o‘Œl

"

‘ ’l“$”$•o–’'“"”$•o–’l“$”$•o–’“”•–’“”•–

’l“$”$•o–’'“"”$•o–’l“$”$•o–’

“”•– —l˜"™išo›—'˜"™$šo›—l˜$™$šo›—˜™šo›—˜™šo›

—l˜"™išo›—'˜"™$šo›—l˜$™$šo›—

˜™šo› œžiŸi o¡œŽiŸ$ o¡œžvŸ$ o¡œžŸi o¡œžŸ$ o¡

œžiŸi o¡œŽiŸ$ o¡œžvŸ$ o¡œŽ

Ÿ$ o¡

§£ ¤ 0

§£ ¤ 0

§£ ¤ 0

§ £ ¤ 0

§ £ ¤ 0

§£ ¤ 0

§£ ¤ 0

§£ ¤ 0

§ £ ¤ 0

¨¦ ¤ 0

¨¦ ¤ 0

¨¦ ¤ 0

¨¦

¤ 0

¨¦ ¤ 0

¨¦ ¤ 0

¨¦ ¤ 0

¨¦ ¤ 0

¨¦

¤ 0

©£ ¤ 0

©£ ¤ 0

©£ ¤ 0

© £ ¤ 0

© £ ¤ 0

©£ ¤ 0

©£ ¤ 0

©£ ¤ 0

© £ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ 0

ª¦ ¤ «

ª¦ ¤ «

ª¦ ¤ «

ª¦ ¤ «

ª¦ ¤ «

ª¦ ¤ «

ª¦

¤ «

ª¦ ¤ «

ª¦

¤ «

© £ ¤ «

© £ ¤ «

© £ ¤ «

©£ ¤ «

©£ ¤ «

© £ ¤ «

© £ ¤ «

© £ ¤ «

©£ ¤ «

¨¦ ¤ «

¨¦ ¤ «

¨¦ ¤ «

¨¦ ¤ «

¨¦ ¤ «

¨¦ ¤ «

¨¦ ¤ «

¨¦ ¤ «

¨¦ ¤ «

§£ ¤ «

§£ ¤ «

§£ ¤ «

§ £ ¤ «

§ £ ¤ «

§£ ¤ «

§£ ¤ «

§£ ¤ «

§ £ ¤ «

¢£ ¤ «

¢£ ¤ «

¢£ ¤ «

¢ £ ¤ «

¢ £ ¤ «

¢£ ¤ «

¢£ ¤ «

¢£ ¤ «

¢ £ ¤ « ® ­ «

® ­ «

® ­ «

® ­ «

® ­ «

®

­ «

®

­ «

®

­ «

® ­ «

¯ ° ±

­

¯ ° ±

­

¯ ° ±

­

¯ ° ± ­

¯ ° ± ­

¯ ° ± ­

¯ ° ± ­

¯ ° ± ­

¯ ° ± ­ ² ³ ´ . 1

² ³ ´ . 1

² ³ ´ . 1

² ³ ´ . 1

² ³ ´ . 1

² ³ ´ . 1

² ³ ´ . 1

² ³ ´ . 1

² ³ ´ . 1

µµ ³.¶

µµ ³.¶

µµ ³.¶

µµ ³. ¶

µµ ³. ¶

µµ ³.¶

µµ ³.¶

µµ ³.¶

µµ ³. ¶ ·¸¹ º

¹

·¸¹ º

¹

·¸¹ º

¹

·¸¹ º

¹

·¸¹ º

¹

·¸¹ º

¹

·¸¹ º

¹

·¸¹ º

¹

·¸¹ º

¹

»¼ ½¾¿ Àº

Á¹

»¼ ½¾¿ ÀºÁ¹

»¼ ½¾¿ À

ºÁ¹

»¼ ½¾¿ Àº

Á¹

»¼ ½¾¿ À

ºÁ¹

»¼ ½¾¿ Àº

Á¹

»¼ ½¾¿ ÀºÁ¹

»¼ ½¾¿ À

ºÁ¹

»¼ ½¾¿ ÀºÁ¹

Figure 2.3: Overview of Locations, where Measurements of Reactive Halogen Species in the troposphere have been reported:

yellow symbols indicate reactive bromine (mostly BrO), green symbols reactive iodine (dark: IO, light: OIO) compounds. GOME satellite measurements are indicated as red hashed areas. The red symbols mark the locations, where measurements of halogen oxides were performed in this work: Alert (BrO, IO), Kuujjuarapik (BrO, upper limits for IO, OIO, OBrO, OClO) and Kerguelen (IO). The light red symbols represent upper limits derived in this work.